Abstract
This report concentrates on specific requirements and goals of the research networks supported by ANSNET, but applies to any TCP dominated high speed WAN and in particular those striving to support high speed end-to-end flows. Measurements have been made under conditions intended to better understand performance barriers imposed by network equipment queueing capacities and queue drop strategies.The IBM RS/6000 based routers currently supporting ANSNET performed very well in these tests. Measurements have been made with the current software and performance enhanced software. Single TCP flows are able to achieve 40 Mb/s and competing multiple TCP flows achieve over 41 Mb/s link utilization on 44.7 Mb/s DS3 links with delays comparable to US cross continent ANSNET delays. Congestion collapse is demonstrated with intentionally reduced queueing capacity and using window sizes much larger than optimal.A variation of Floyd and Jacobson's Random Early Detection (RED) algorithm [1] is tested. Performance improved with the use of RED for tests involving multiple flows. With RED and queueing capacity at or above the delay bandwidth product, congestion collapse is avoided, allowing the maximum window size to safely be set arbitrarily high.Queueing capacity greater than or equal to the delay bandwidth product and RED are recommended. RED provides performance improvement in all but the single flow case, but cannot substitute for adequate queueing capacity, particularly if high speed flows are to be supported.
- {1} S. Floyd and V. Jacobson, "Random early detection gateways for congestion avoidance," IEEE/ACM Transactions on Networking, vol. 1, pp. 397-413, Aug. 1993. Google ScholarDigital Library
- {2} G. T. Almes, "Engineering trans-oceanic 44 mb/s tcp/ip networks." presented at the Third Symposium on High Speed Networking for Research in Europe, Brussels, Feb 1994.Google Scholar
- {3} K. C. Claffy, H.-W. Braun, and G. C. Polyzos, "Long-term traffic aspects of the NSFNET," in Proceedings of the International Networking Conference INET), (San Francisco, California), pp. CBA.-1- CBA-12, Internet Society, Aug. 1993.Google Scholar
- {4} ANS Network Operations Center, "Ansnet link utilization daily reports." unpublished reports used for capacity planning, Oct 1994.Google Scholar
- {5} ANS Network Operations Center, "Ansnet congestion loss daily reports." unpublished reports used for capacity planning, Oct 1994.Google Scholar
- {6} J. Nagle, "Congestion control in IP/TCP internetworks," ACM Computer Communication Review, vol. 14, pp. 11-17, Oct. 1984. Google ScholarDigital Library
- {7} V. Jacobson, "Congestion avoidance and control," ACM Computer Communication Review, vol. 18, pp. 314-329, Aug. 1988. Proceedings of the Sigcomm '88 Symposium in Stanford, CA, August, 1988. Google ScholarDigital Library
- {8} J. Postel, "Transmission control protocol," Network Working Group Request for Comments RFC 793, Information Sciences Institute, University of Southern California, Sept. 1981.Google Scholar
- {9} D. Clark, "Window and acknowlegement strategy in TCP," Request for Comments RFC 813, Internet Engineering Task Force, July 1982.Google Scholar
- {10} J. Nagle, "Congestion control in IP/TCP internetworks," Request for Comments RFC 896, Internet Engineering Task Force, Jan. 1984.Google Scholar
- {11} J. Postel, "TCP maximum segment size and related topics," Request for Comments RFC 879, Internet Engineering Task Force, Nov. 1983.Google Scholar
- {12} D. Mills, "Internet delay experiments," Request for Comments RFC 889, Internet Engineering Task Force, Dec. 1983.Google Scholar
- {13} R. Braden, "Requirements for internet hosts - communication layers," Network Working Group Request for Comments RFC 1122, Information Sciences Institute, Oct. 1989.Google Scholar
- {14} L. Peterson and S. O'Malley, "TCP extensions considered harmful," Request for Comments (Informational) RFC 1263, Internet Engineering Task Force, Oct. 1991.Google Scholar
- {15} R. Braden and V. Jacobson, "TCP extensions for long-delay paths," Request for Comments (Experimental) RFC 1072, Internet Engineering Task Force, Oct. 1988.Google Scholar
- {16} R. Fox, "TCP big window and NAK options," Request for Comments RFC 1106, Internet Engineering Task Force, June 1989.Google Scholar
- {17} A. McKenzie, "Problem with the TCP big window option," Request for Comments RFC 1110, Internet Engineering Task Force, Aug. 1989.Google Scholar
- {18} R. Braden, V. Jacobson, and L. Zhang, "TCP extension for High-Speed paths," Request for Comments (Experimental) RFC 1185, Internet Engineering Task Force, Oct. 1990.Google Scholar
- {19} D. Borman, R. Braden, and V. Jacobson, "TCP extensions for high performance," Request for Comments (Proposed Standard) RFC 1323, Internet Engineering Task Force, May 1992. Obsoletes RFC1185.Google Scholar
- {20} W. R. Stevens, TCP/IP illustrated: the protocols, vol. 1. Reading, Massachusetts : Addison-Wesley, 1994. Google ScholarDigital Library
- {21} D. Clark, V. Jacobson, J. Romkey, and M. Salwen, "An analysis of TCP processing overhead," IEEE Communications Magazine, vol. 27, pp. 23-29, June 1989.Google ScholarDigital Library
- {22} D. A. Borman, "Implementing TCP/IP on a cray computer," ACM Computer Communication Re - view, vol. 19, pp. 11-15, Apr. 1989. Google ScholarDigital Library
- {23} V. Jacobson, "Some design issues for high-speed networks," in Networkshop '93, (Melbourne, Australia), Nov. 1993.Google Scholar
- {24} J. Mogul and S. Deering, "Path MTU discovery," Request for Comments (Draft Standard) RFC 1191, Internet Engineering Task Force, Nov. 1990.Google Scholar
- {25} S. Knowles, "IESG advice from experience with path MTU discovery," Request for Comments (Informational) RFC 1435, Internet Engineering Task Force, Mar. 1993.Google Scholar
- {26} S. Floyd and V. Jacobson, "On traffic phase effects in packet-switched gateways," Internetworking : Research and Experience, vol. 3, pp. 115-156, Sept. 1992.Google Scholar
- {27} S. Shenker, L. Zhang, and D. Clark, "Some observations on the dynamics of a congestion control algorithm," ACM Computer Communication Review, vol. 20, pp. 30-39, Oct. 1990. Google ScholarDigital Library
- {28} L. Zhang, S. Shenker, and D. D. Clark, "Observations on the dynamics of a congestion control algorithm : the effects of two-way traffic," in Sigcomm '91 Conference: Communications Architectures and Protocols, (Zürich, Switzerland), pp. 133- 147, ACM, Sept. 1991. Google ScholarDigital Library
- {29} K. C. Claffy, H.-W. Braun, and G. C. Polyzos, "Tracking long-term growth of the NSFNET," Communications ACM, vol. 37, pp. 34-45, 1994. Google ScholarDigital Library
- {30} C. Villamizar, "A variation of random early detection congestion avoidance." work in progress.Google Scholar
- {31} R. Cáceres, P. B. Danzig, S. Jamin, and D. J. Mitzel, "Characteristics of wide-area TCP/IP conversations," in SIGCOMM Symposium on Communications Architectures and Protocols, (Zürich, Switzerland), pp. 101-112, ACM, Sept. 1991. also in Computer Communication Review 21 (4), Sep. 1991. Google ScholarDigital Library
- {32} V. Paxson, "Empirically-derived analytic models of wide-area TCP connections," technical report, Lawrence Berkeley Laboratory and EECS Division, University of California, Berkeley, California, June 1993.Google Scholar
- {33} W. E. Leland, M. S. Taqq, W. Willinger, and D. V. Wilson, "On the self-similar nature of Ethernet traffic," in SIGCOMM Symposium on Communications Architectures and Protocols (D. P. Sidhu, ed.), (San Francisco, California), pp. 183-193, ACM, Sept. 1993. also in Computer Communication Review 23 (4), Oct. 1992. Google ScholarDigital Library
- {34} V. Paxson and S. Floyd, "Wide-area traffic: the failure of Poisson modeling," in SIGCOMM Symposium on Communications Architectures and Protocols , (London, United Kingdom), pp. -, ACM, Aug. 1994. Google ScholarDigital Library
- {35} C. Partridge, "Editorial : The end of simple traffic models," IEEE Network, vol. 7, Sept. 1993.Google Scholar
- {36} T. Skibo, "Experiences with tcp extensions for high-performance networks." contained in ftp://vixen.cso.uiuc.edu/pub/tcplw.shar.Z.Google Scholar
- {37} S. Floyd, "Connections with multiple congested gateways in packet-switched networks part 1 : One-way traffic," ACM Computer Communication Review , vol. 21, pp. 30-47, Oct. 1991. Google ScholarDigital Library
- {38} I. Merit, "Nsfnet statistics repository." ftp://nis.nsf.net/statistics/nsfnet/.Google Scholar
- {39} A. Romanow and S. Floyd, "Dynamics of TCP traffic over ATM networks." see also 6th IEEE LAN/MAN Workshop, 1993.Google Scholar
Index Terms
- High performance TCP in ANSNET
Recommendations
Sync-TCP in high-speed environments
ACM-SE 43: Proceedings of the 43rd annual Southeast regional conference - Volume 2The congestion control mechanism in TCP was first introduced by Jacobson in [7], which was later developed into TCP Tahoe. Since then, various end-to-end congestion control protocols have been proposed, including Reno [2], NewReno [6], SACK [4], and ...
TCP CERL: congestion control enhancement over wireless networks
In this paper, we propose and verify a modified version of TCP Reno that we call TCP Congestion Control Enhancement for Random Loss (CERL). We compare the performance of TCP CERL, using simulations conducted in ns-2, to the following other TCP variants: ...
Performance evaluation of TCP-BIAD in high-speed, long-distance networks
In this paper, the performance of Binary Increase Adaptive Decrease (TCP-BIAD) congestion control algorithm in high-speed long-distance networks is evaluated. As its name implies, this TCP variant is a combination of an enhanced binary increase ...
Comments